NITANSH THAKERNT / SYSTEMS

BSc General Engineering — Cyber Systems at DTU

Nitansh Thaker

Autonomous systems, from sensed state to verified action.

Mission control, reinforcement learning, and embedded software for intelligent physical systems.

Explore selected work Download CV
Autonomous-system sense-to-verification assemblyA diagrammatic autonomous-system model, not deployed hardware. A mounted controller sends an amber command through an external cable to a sensor head on a two-joint arm; pale-blue encoder feedback returns from the shoulder joint to the controller.CTRL01SENSESTATE ACQUISITION02DECIDECONTROLLER +POLICY03COMMUNICATECOMMAND + ACK04ACTJOINT MOTORS +PHYSICAL OUTPUT05VERIFYENCODER FEEDBACK
FIG. A · AUTONOMY ASSEMBLYDIAGRAMMATIC SYSTEM MODEL / NOT DEPLOYED HARDWARE
  1. 01SENSEAcquire system stateSensors · simulation
  2. 02DECIDEChoose within constraintsState machine · policy
  3. 03COMMUNICATESend command and contextBinary · messaging
  4. 04ACTChange system stateTransition · dispatch
  5. 05VERIFYObserve the resultACK · tests · traces

Systems, decisions, evidence.

Two primary case studies show the same engineering habit in different operating environments: make state explicit, constrain decisions, and keep verification close to the action.

Mission Control Engineer
DanSTAR Rocketry

Mission Control

Mission-control and flight-computer software built around explicit state, validated commands, and acknowledged transitions.

DanSTAR command validation and acknowledgement architectureA diagrammatic architectural example: Mission Control validates and sends an ARMED to ASCENT request to the Flight Computer. The requested state remains pending until an acknowledgement returns to Mission Control.ARCHITECTURAL EXAMPLE · ARMED → ASCENTVALIDATEpayload + eligibilitySENDto flight computerAWAIT ACKtransition pendingRETURN ACKback to controlCONFIRMcommit after ACKVALIDATED COMMAND →SOURCE + VALIDATORMISSION CONTROLvalidate payloadCOMMAND RECEIVERFLIGHT COMPUTERreceive + ACK← ACKNOWLEDGEMENT RETURNSCURRENT STATEARMEDREQUESTED STATEASCENTNOT SENTCONFIRMED STATE—
Command path / state verificationSystem abstraction / verified project logic
Context
DanSTAR Rocketry · Technical University of Denmark
System
A flowchart-aligned mission state machine with phase-specific Continue, Abort, and Reset logic.
Decision
Treat flight-computer acknowledgement as the boundary for a confirmed transition.
Boundary
The portfolio describes software architecture and verification practices, not flight results.
Read the case study
Research Intern
PHAETHON Centre of Excellence

Reinforcement Learning for Local Energy Markets

Simulation and evaluation of constrained battery and market decisions across transparent baselines and learning-based control.

PHAETHON reinforcement-learning decision loopA diagrammatic reinforcement-learning decision loop for PHAETHON local-energy-market research. Current observations and separate policy constraints inform a constrained policy, which selects one eligible action. The battery and grid environment processes that action, produces a system response, and updates the next observation for the following decision step.OBSERVATIONSCURRENT STATEPV GENERATIONDEMAND / LOADMARKET SIGNALBATTERY SOCCONSTRAINTSSOC BOUNDSACTIONELIGIBILITYCONSTRAINEDPOLICYELIGIBLEACTIONSYSTEM /ENVIRONMENTBATTERY + GRIDSYSTEMRESPONSESTATE TRANSITIONNEXT STATEUPDATEDOBSERVATION
Scroll to inspect the full diagram →
State / policy / constrained actionSystem abstraction / verified project logic
Context
PHAETHON Centre of Excellence · Cyprus
System
A Python environment modelling photovoltaic generation, demand, battery storage, grid import/export, and time-varying prices.
Decision
Retain a transparent rule-based policy as a behavioural and economic reference.
Boundary
The work is simulation and evaluation, not deployment on a physical robot or live market.
Read the case study

Autonomy is a chain of accountable transitions.

I am interested in systems where sensing quality, state representation, communications, control decisions, and safe operation cannot be separated. The portfolio uses that chain as its reading order—not as simulated instrumentation.

Current capability

Mission state machines, command validation, simulation, reinforcement-learning environments, and embedded integration.

Direction of travel

Autonomous systems, learning-enabled robotics, and reliable software around intelligent physical systems.

Working standard

Distinguish what was built, what was tested, what remains exploratory, and what evidence is not public.

A wider system index.

Supporting projects move through physical sensing, web-connected monitoring, energy dispatch, program execution, and visual perception.

DTU Embedded Systems Programming

Sports Timing and Control System

An ESP32 prototype integrating sensing, state feedback, physical actuation, displays, and race-control logic.

Embedded sports timer sensing, timing-state, and feedback loop01READY02COUNTDOWN03ARMED04RUNNING05FINISHEDULTRASONIC EVENT · TIMER · MULTI-MODAL FEEDBACK
Scroll to inspect the full diagram →
Open case study
DTU Design-Build 4

Biological-System Monitoring

ESP32 sensing and a web interface for monitoring a living mussel system under biological operating constraints.

Biological monitoring with two temperature channels, one light channel, and a safe operating rangeBIOLOGICAL OPERATING RANGETEMPERATURE 01TEMPERATURE 02LIGHT
Scroll to inspect the full diagram →
Open case study
Independent engineering project

Limassol Hotel Battery Dispatch

A testable Django service for deterministic 15-minute photovoltaic, load, tariff, and battery dispatch reporting.

Battery dispatch sequence from solar to load, storage, and gridA diagrammatic model of an explainable greedy policy operating at 15-minute intervals. Solar generation passes to Load, then Battery and Grid stages. Each box fills from left to right to represent processing, and one cobalt circle represents transfer between completed stages.01SOLAR02LOAD03BATTERY04GRID15-MINUTE INTERVAL · EXPLAINABLE GREEDY POLICY
Scroll to inspect the full diagram →
Open case study
DTU Cyber Systems

Instruction-Set Architecture Simulator

A modular Python simulator for registers, memory, parsing, arithmetic, logic, and branching instructions.

Instruction-set simulator register and execution viewPC0x1000R00x2100R10x3200R20x4300MEM0x5400CURRENT INSTRUCTIONBRANCH R1, +04PARSE → EXECUTE → OBSERVE
Scroll to inspect the full diagram →
Open case study
Independent project

Automated Wordle Solver with Visual Perception

A real-time assistant combining board detection, OCR, colour classification, and heuristic action selection.

Visual-perception pipeline from screenshot to extracted state and recommendationA Wordle-recognition pipeline. A cobalt-blue line scans the grid for OCR. Green tiles mean a letter is correct and in the correct position; amber tiles mean a letter is present in the wrong position; neutral tiles are unknown or absent. The interpreted colour state informs a heuristic action.SCAN / OCRCRANESHEARPRONEOCR + COLOURSTATEHEURISTICACTION
Open case study

From physical constraints to intelligent action.

My path runs through Cyprus, Denmark, and an exchange semester at HKUST—across embedded programming, mission operations, energy research, and the broader study of autonomous systems.

View experience About the direction
Geographic trajectory from Cyprus to Denmark to Hong KongA cropped geographic map of Europe, Africa, Asia, and Oceania. Cyprus, Copenhagen in Denmark, and Hong Kong are plotted using geographic coordinates. The route represents a biographical and academic trajectory from Cyprus to Copenhagen and then Hong Kong.01CYPRUS02DENMARK03HONG KONG

HOVER / TAP TAGS FOR DETAILS

CYPRUS → DENMARK → HONG KONG
  1. 01
    CYPRUSORIGIN AND EARLY PROJECTS
  2. 02
    DENMARKCOPENHAGEN · DTU AND DANSTAR
  3. 03
    HONG KONGHKUST EXCHANGE · AUTUMN 2026